Github user JoshRosen commented on a diff in the pull request:
https://github.com/apache/spark/pull/2490#discussion_r22244391
--- Diff: docs/programming-guide.md ---
@@ -1183,6 +1189,10 @@ running on the cluster can then add to it using the
`add` method or the `+=` ope
However, they cannot read its value.
Only the driver program can read the accumulator's value, using its
`value` method.
+An operation referencing an accumulator may run multiple times, either
when parts of its output
+data become lost or when multiple actions make use of the RDD it produces.
In these cases, only the
+additions reported by the first successful execution contribute to the
accumulator's value.
--- End diff --
I was going to suggest that this might not be true, but it looks like this
is partially-true now that #2524 has been merged. That PR added a paragraph to
the programming guide which clarifies that we guard against duplicate updates
only for updates performed inside of actions and _not_ for ones performed in
transformations:
https://github.com/apache/spark/commit/66cc2431462a5354bb50c196a59da0ffc258c466?diff=unified#diff-3
In light of this, do we still need this paragraph?
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
